A geologic model is a computer-based representation of a region of the earth subsurface. Such models are typically used to model a petroleum reservoir or a depositional basin. A geologic model commonly comprises a three dimensional (3-D) geocellular grid that is composed of contiguous 3-D cells. Each of the cells is assigned various properties, such as lithology, porosity, permeability, and/or water saturation, using various algorithms, e.g., geostatistical algorithms. After formation, the geologic model can be used for many purposes. One common use for the geologic model is as an input to a computer program that simulates the movement of fluids within the modeled subsurface region. These types of programs are used to predict, for example, hydrocarbon production rates and volumes from a petroleum reservoir over time.
Despite the usefulness of this technology, the current applications of the technology have several problems. For instance, one problem is that the geologic models do not precisely represent the geologic description of the region of interest. That is, current technology is not able to construct geologic models that precisely represent the characteristics of the interpreted or conceived geologic description which have a significant effect on the movement of fluids in the reservoir. Characteristics can include, for example, the compositions, dimensions, geometries, orientations, locations, and spatial, topological and hierarchical associations of various descriptive elements. The geologic description may also include information on spatial trends and/or changes in these elements, e.g., trends in composition, dimension and orientation. The descriptive elements represent regions of any scale within the reservoir, and the boundaries that separate contiguous regions. Regions can include but are not limited to stratigraphic regions, such as sequences or parasequences, facies regions, such as shale layers or individual channel facies, diagenetic regions, such as cemented or porous regions and fractured regions, and structural regions, such as fault blocks separated by fault planes. Geologic models built using current technology are not able to precisely represent the geologic characteristics that effect fluid flow within the model, because the technology was initially developed from subsurface mining, which was less concerned about fluid movement. As such, reliance on these model-based predictions are problematic when they are used as a basis for making business decisions, such as decisions relating to drilling and completing wells, and to constructing surface facilities to handle the production of hydrocarbons.
The geologic description used may be based on different techniques. For instance, an interpreted description is one that is derived by analyzing data obtained from the subsurface region being modeled. Alternatively, a conceived description is not or cannot be directly analyzed from these subsurface data, but is assumed to be accurate based on analog data and individual experience.
Inaccurate geologic description arises from one or more different factors. One factor is that the various descriptive geologic elements differ significantly in scale, but only a narrow range of scales can be precisely represented in the geologic model. In part, this is because all cells that constitute most geocellular grids have similar dimensions; i.e., represent approximately a single scale. As a result, the model can not explicitly represent descriptive geologic elements having scales finer than the grid cell dimensions commonly used in known simulation programs.
Another factor is that stochastic geologic modeling algorithms commonly used to form the geologic model are limited in their ability to precisely represent the descriptive elements, particularly if these elements are in minor abundance. Geostatistical simulation algorithms cannot reproduce geologic elements having long-range spatial correlation, such as facies having geometries represented by large sheets or long channels. Object-based algorithms (also referred to as Marked-Point or Boolean algorithms) can produce sheet and channel facies elements, yet one can not control placement of these facies in the model. For example, such control is desirable to control connectivity between well locations. As such, the modeling algorithms further limit the geologic description.
Yet another factor is that the geologic model, which is directly used for simulating fluid movement in the reservoir (e.g., the simulation model), may have limited resemblance to the input geologic description. Because of specific grid requirements for flow simulation, the simulation model is often constructed on a different grid from that of the initial geologic model. That is, to efficiently perform the simulation of fluid flow, the initial geologic model has its grid cells and their properties coarsened, which results in scale-averaging of the properties within the model. This re-gridding and scale-averaging can result in further distortions of the input geologic descriptions.
A second problem associated with current geologic modeling technology is an inability to rapidly construct and update the geologic model. Many modeling workflows require that multiple geologic models be constructed or updated, such as workflows associated with analyzing for the effects of uncertainty on flow predictions, or with optimizing the model to match field production data, e.g., history matching. For instance, as the geologic models may be large, the current models are difficult and inefficient to update to include new data from a new well. Inefficiency itself can lead to increased cost, though greatest risk comes when time constraints either prohibit these workflows or limit their effectiveness. Additionally, the geologic models often contain an abundance of geologic detail that is unnecessary to accurately simulate fluid-flow behavior in the model.
Considerable effort and time is required to construct and update such a geologic model, and this extra management of the geologic model further limits the construction and updates of these models.
Other related material may be found in U.S. Pat. Nos. 5,905,657; 6,035,255; 6,044,328; and 6,191,787.    Dubrule, O., et al., 1997, Reservoir Geology Using 3-D Modeling Tools, SPE 38659.    Landis, Lester H. and Peter N. Glenton, 2007, Reservoir Model Building Methods, published U.S. patent application 2007/0061117.    Murphy, William F. et al., 2000c, Apparatus for Creating, Testing, and Modifying Geological Subsurface Models, U.S. Pat. No. 6,070,125.    Pyrcz, M. J., et al., 2005, Stochastic Surface-Based Modeling of Turbidite Lobes, AAAPG Bulletin, V. 89, No. 2, pp. 177-191.    Scaglioni, P. et al., 2006, Implicit Net-to-Gross in the Petrophysical Characterization of Thin-Layered Reservoirs, Petroleum Geoscience, V. 12, pp. 325-333.    Wen, w., et al., 1998, Three-Dimensional Simulation of Small-Scale Heterogeneity in Tidal Deposits—a Process-Based Stochastic Simulation Method. In: Buccianti, A. et al., (eds.), Proceedings of the 4th Annual Conference of the International Association of Mathematical Geology (IAMG), Naples, pp. 129-134.    Wentland, Robert and Peter Whitehead, 2007a, Pattern Recognition Template Construction Applied to Oil Exploration and Production, U.S. Pat. No. 7,162,463.    Wentland, Robert and Peter Whitehead, 2007b, Pattern Recognition Template Application Applied to Oil Exploration and Production, U.S. Pat. No. 7,188,092.